Merge 4b1c830957 into 4d05f84325

* no space after @ in decorator

test/test_jsinterp.py

@@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
     def test_unary_operators(self):
         jsi = JSInterpreter('function f(){return 2  -  - - 2;}')
         self.assertEqual(jsi.call_function('f'), 0)
-        # fails
+        jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
-        # self.assertEqual(jsi.call_function('f'), 0)
+        # https://github.com/ytdl-org/youtube-dl/issues/32815
+        jsi = JSInterpreter('function f(){return 0  - 7 * - 6;}')
+        self.assertEqual(jsi.call_function('f'), 42)
 
     """ # fails so far
     def test_packed(self):

test/test_youtube_signature.py

@@ -158,6 +158,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
     ),
+    (
+        'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
+        '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
+    ),
 ]
 
 

youtube_dl/extractor/common.py

@@ -3033,7 +3033,6 @@ class InfoExtractor(object):
             transform_source=transform_source, default=None)
 
     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
-
         # allow passing `transform_source` through to _find_jwplayer_data()
         transform_source = kwargs.pop('transform_source', None)
         kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

youtube_dl/extractor/extractors.py

@@ -1069,7 +1069,7 @@ from .rtve import RTVEALaCartaIE, RTVELiveIE, RTVEInfantilIE, RTVELiveIE, RTVETe
 from .rtvnh import RTVNHIE
 from .rtvs import RTVSIE
 from .ruhd import RUHDIE
-from .rumble import RumbleEmbedIE
+from .rumble import RumbleIE
 from .rutube import (
     RutubeIE,
     RutubeChannelIE,
@@ -1439,6 +1439,7 @@ from .urort import UrortIE
 from .urplay import URPlayIE
 from .usanetwork import USANetworkIE
 from .usatoday import USATodayIE
+from .usawatchdog import UsaWatchdogIE
 from .ustream import UstreamIE, UstreamChannelIE
 from .ustudio import (
     UstudioIE,

youtube_dl/extractor/generic.py

@@ -132,6 +132,7 @@ from .kinja import KinjaEmbedIE
 from .arcpublishing import ArcPublishingIE
 from .medialaan import MedialaanIE
 from .simplecast import SimplecastIE
+from .rumble import RumbleIE
 
 
 class GenericIE(InfoExtractor):
@@ -3518,6 +3519,10 @@ class GenericIE(InfoExtractor):
             return self.playlist_from_matches(
                 zype_urls, video_id, video_title, ie=ZypeIE.ie_key())
 
+        rumble_urls = RumbleIE.rumble_embedded_id(webpage)
+        if rumble_urls is not None:
+            return self.playlist_result(rumble_urls) if len(rumble_urls) > 1 else rumble_urls[0]
+
         # Look for HTML5 media
         entries = self._parse_html5_media_entries(url, webpage, video_id, m3u8_id='hls')
         if entries:

youtube_dl/extractor/palcomp3.py

@@ -8,7 +8,7 @@ from ..compat import compat_str
 from ..utils import (
     int_or_none,
     str_or_none,
-    try_get,
+    traverse_obj,
 )
 
 
@@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
     }
     name'''
 
-    @ classmethod
+    @classmethod
     def suitable(cls, url):
         return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
 
@@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
         artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
 
         def entries():
-            for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []):
+            for music in traverse_obj(artist, (
+                    'musics', 'nodes', lambda _, m: m['musicID'])):
                 yield self._parse_music(music)
 
         return self.playlist_result(
@@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
             'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
             'description': 'md5:7043342c09a224598e93546e98e49282',
             'upload_date': '20161107',
-            'uploader_id': 'maiaramaraisaoficial',
+            'uploader_id': '@maiaramaraisaoficial',
             'uploader': 'Maiara e Maraisa',
         }
     }]

youtube_dl/extractor/rumble.py

@@ -1,5 +1,6 @@
 # coding: utf-8
 from __future__ import unicode_literals
+import re
 
 from .common import InfoExtractor
 from ..compat import compat_str
@@ -8,31 +9,100 @@ from ..utils import (
     int_or_none,
     parse_iso8601,
     try_get,
+    ExtractorError,
 )
 
 
-class RumbleEmbedIE(InfoExtractor):
+class RumbleIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?rumble\.com/embed/(?:[0-9a-z]+\.)?(?P<id>[0-9a-z]+)'
-    _TESTS = [{
-        'url': 'https://rumble.com/embed/v5pv5f',
-        'md5': '36a18a049856720189f30977ccbb2c34',
-        'info_dict': {
-            'id': 'v5pv5f',
-            'ext': 'mp4',
-            'title': 'WMAR 2 News Latest Headlines | October 20, 6pm',
-            'timestamp': 1571611968,
-            'upload_date': '20191020',
-        }
-    }, {
-        'url': 'https://rumble.com/embed/ufe9n.v5pv5f',
-        'only_matching': True,
-    }]
 
-    def _real_extract(self, url):
+    RE_DICT = {
-        video_id = self._match_id(url)
+        'iframe_url': {
+            're': r'https?://(?:www\.)?rumble\.com/embed/(?:[0-9a-z]+\.)?(?P<id>[0-9a-z]+)',
+            'compiled': None},
+        'jscript_url': {
+            're': r'https?://rumble\.com/[a-zA-Z0-9-_.]*\.html',
+            'compiled': None},
+        'list_url': {
+            're': r'https?://rumble.com/(?:c|user)/(?P<id>[^/]+)',
+            'compiled': None},
+        'jscript_id': {
+            're': r'Rumble *\( *["\']play["\'], *\{[^}]*["\']video["\'] *: *["\'](?P<id>[^"\']+)',
+            'compiled': None}
+    }
+
+    _TESTS = [
+        {
+            'url': 'https://rumble.com/embed/v5pv5f',
+            'md5': '36a18a049856720189f30977ccbb2c34',
+            'info_dict': {
+                'id': 'v5pv5f',
+                'ext': 'mp4',
+                'title': 'WMAR 2 News Latest Headlines | October 20, 6pm',
+                'timestamp': 1571611968,
+                'upload_date': '20191020',
+            }
+        },
+        {
+            'url': 'https://rumble.com/v8c1bt-wmar-2-news-latest-headlines-october-20-6pm.html',
+            'md5': '36a18a049856720189f30977ccbb2c34',
+            'info_dict': {
+                'id': 'v5pv5f',
+                'ext': 'mp4',
+                'title': 'WMAR 2 News Latest Headlines | October 20, 6pm',
+                'timestamp': 1571611968,
+                'upload_date': '20191020',
+            }
+        },
+        {
+            'url': 'https://rumble.com/c/PeakProsperity',
+            'playlist_mincount': 25,
+            'info_dict': {
+                'id': 'PeakProsperity',
+            }
+        },
+        {
+            'url': 'https://rumble.com/embed/ufe9n.v5pv5f',
+            'only_matching': True,
+        }
+    ]
+
+    @classmethod
+    def get_re(cls, tag):
+        if cls.RE_DICT[tag]['compiled'] is None:
+            cls.RE_DICT[tag]['compiled'] = re.compile(cls.RE_DICT[tag]['re'])
+        return cls.RE_DICT[tag]['compiled']
+
+    @classmethod
+    def suitable(cls, url):
+        return (cls.get_re('jscript_url').match(url) is not None
+                or cls.get_re('list_url').match(url) is not None
+                or cls.get_re('iframe_url').match(url) is not None)
+
+    @staticmethod
+    def rumble_embedded_id(page_data):
+        '''For use by extractors of sites which use emedded Rumble videos. Given
+        a webpage as a string returns a list of url result dicts for each embedded
+        rumble video found. None is returned if no embeds were found. Duplicates
+        are not removed'''
+
+        embeds = []
+        # The JS embeds
+        for mobj in RumbleIE.get_re('jscript_id').finditer(page_data):
+            embeds.append(InfoExtractor.url_result('https://rumble.com/embed/' + mobj.group('id'), 'Rumble', mobj.group('id')))
+
+        # The iframes embeds
+        for mobj in RumbleIE.get_re('iframe_url').finditer(page_data):
+            embeds.append(InfoExtractor.url_result('https://rumble.com/embed/' + mobj.group('id'), 'Rumble', mobj.group('id')))
+
+        return embeds if embeds else None
+
+    def rumble_video_info(self, video_id):
         video = self._download_json(
             'https://rumble.com/embedJS/', video_id,
             query={'request': 'video', 'v': video_id})
+        if not video:
+            raise ExtractorError('Unable to locate video information.', expected=True)
+
         title = video['title']
 
         formats = []
@@ -65,3 +135,23 @@ class RumbleEmbedIE(InfoExtractor):
             'channel_url': author.get('url'),
             'duration': int_or_none(video.get('duration')),
         }
+
+    def _real_extract(self, url):
+        if self.get_re('jscript_url').match(url) is not None:
+            page = self._download_webpage(url, 'Rumble Page')
+            video_id = self._search_regex(self.get_re('jscript_id'), page, "id")
+            return self.rumble_video_info(video_id)
+
+        mobj = self.get_re('list_url').match(url)
+        if mobj is not None:
+            urls = []
+            id = mobj.group('id')
+            page = self._download_webpage(url, id)
+            for mobj in re.finditer(r'<a class=video-item--a href=\/(?P<href>[a-zA-Z0-9\-.]+)>', page):
+                urls.append('https://rumble.com/' + mobj.group('href'))
+
+            return self.playlist_from_matches(urls, id)
+
+        mobj = self.get_re('iframe_url').match(url)
+        if mobj is not None:
+            return self.rumble_video_info(mobj.group('id'))

youtube_dl/extractor/usawatchdog.py

@@ -0,0 +1,24 @@
+# -*- coding: utf-8 -*-
+from __future__ import unicode_literals
+import re
+
+from .common import InfoExtractor
+
+
+class UsaWatchdogIE(InfoExtractor):
+    _VALID_URL = r'^https?://(?:www\.)?usawatchdog\.com/?$'
+    _TEST = {
+        'url': 'https://usawatchdog.com/',
+        'playlist_mincount': 15,
+        'info_dict': {
+            'id': 'USA Watchdog',
+        }}
+
+    def _real_extract(self, url):
+        matches = []
+        for mobj in re.finditer(r'front-view-title[^<]+<a.+href=["\'](?P<href>https?:(?:www\.)?//usawatchdog.com/[^/]+\/?)[^>]+>(?P<title>[^<]+)',
+                                self._download_webpage(url, 'Site Root')):
+            matches.append(self.url_result(mobj.group('href'),
+                                           video_title=mobj.group('title').encode('utf8')))
+
+        return self.playlist_result(matches, 'USA Watchdog')

youtube_dl/jsinterp.py

@@ -14,6 +14,7 @@ from .utils import (
     remove_quotes,
     unified_timestamp,
     variadic,
+    write_string,
 )
 from .compat import (
     compat_basestring,
@@ -53,15 +54,16 @@ def wraps_op(op):
 
 # NB In principle NaN cannot be checked by membership.
 # Here all NaN values are actually this one, so _NaN is _NaN,
-# although _NaN != _NaN.
+# although _NaN != _NaN. Ditto Infinity.
 
 _NaN = float('nan')
+_Infinity = float('inf')
 
 
 def _js_bit_op(op):
 
     def zeroise(x):
-        return 0 if x in (None, JS_Undefined, _NaN) else x
+        return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
 
     @wraps_op(op)
     def wrapped(a, b):
@@ -84,7 +86,7 @@ def _js_arith_op(op):
 def _js_div(a, b):
     if JS_Undefined in (a, b) or not (a or b):
         return _NaN
-    return operator.truediv(a or 0, b) if b else float('inf')
+    return operator.truediv(a or 0, b) if b else _Infinity
 
 
 def _js_mod(a, b):
@@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
         return 'LocalNameSpace%s' % (self.maps, )
 
 
+class Debugger(object):
+    ENABLED = False
+
+    @staticmethod
+    def write(*args, **kwargs):
+        level = kwargs.get('level', 100)
+
+        def truncate_string(s, left, right=0):
+            if s is None or len(s) <= left + right:
+                return s
+            return '...'.join((s[:left - 3], s[-right:] if right else ''))
+
+        write_string('[debug] JS: {0}{1}\n'.format(
+            '  ' * (100 - level),
+            ' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
+
+    @classmethod
+    def wrap_interpreter(cls, f):
+        def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
+            if cls.ENABLED and stmt.strip():
+                cls.write(stmt, level=allow_recursion)
+            try:
+                ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
+            except Exception as e:
+                if cls.ENABLED:
+                    if isinstance(e, ExtractorError):
+                        e = e.orig_msg
+                    cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
+                raise
+            if cls.ENABLED and stmt.strip():
+                if should_ret or not repr(ret) == stmt:
+                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+            return ret, should_ret
+        return interpret_statement
+
+
 class JSInterpreter(object):
     __named_object_counter = 0
 
@@ -307,8 +345,7 @@ class JSInterpreter(object):
     def __op_chars(cls):
         op_chars = set(';,[')
         for op in cls._all_operators():
-            for c in op[0]:
+            op_chars.update(op[0])
-                op_chars.add(c)
         return op_chars
 
     def _named_object(self, namespace, obj):
@@ -326,9 +363,8 @@ class JSInterpreter(object):
         # collections.Counter() is ~10% slower in both 2.7 and 3.9
         counters = dict((k, 0) for k in _MATCHING_PARENS.values())
         start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
-        in_quote, escaping, skipping = None, False, 0
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
-        after_op, in_regex_char_group = True, False
+        skipping = 0
-
         for idx, char in enumerate(expr):
             paren_delta = 0
             if not in_quote:
@@ -382,10 +418,12 @@ class JSInterpreter(object):
         return separated[0][1:].strip(), separated[1].strip()
 
     @staticmethod
-    def _all_operators():
+    def _all_operators(_cached=[]):
-        return itertools.chain(
+        if not _cached:
-            # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+            _cached.extend(itertools.chain(
-            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+                # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+                _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
+        return _cached
 
     def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
         if op in ('||', '&&'):
@@ -416,7 +454,7 @@ class JSInterpreter(object):
         except Exception as e:
             if allow_undefined:
                 return JS_Undefined
-            raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e)
+            raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
         try:
@@ -438,6 +476,7 @@ class JSInterpreter(object):
     _FINALLY_RE = re.compile(r'finally\s*\{')
     _SWITCH_RE = re.compile(r'switch\s*\(')
 
+    @Debugger.wrap_interpreter
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
             raise self.Exception('Recursion limit reached')
@@ -511,7 +550,6 @@ class JSInterpreter(object):
                 expr = self._dump(inner, local_vars) + outer
 
         if expr.startswith('('):
-
             m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
             if m:
                 # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@@ -693,7 +731,7 @@ class JSInterpreter(object):
                 (?P<op>{_OPERATOR_RE})?
                 =(?!=)(?P<expr>.*)$
             )|(?P<return>
-                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+                (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
             )|(?P<indexing>
                 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
             )|(?P<attribute>
@@ -727,11 +765,12 @@ class JSInterpreter(object):
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
-
         elif expr == 'undefined':
             return JS_Undefined, should_return
         elif expr == 'NaN':
             return _NaN, should_return
+        elif expr == 'Infinity':
+            return _Infinity, should_return
 
         elif md.get('return'):
             return local_vars[m.group('name')], should_return
@@ -760,18 +799,28 @@ class JSInterpreter(object):
             right_expr = separated.pop()
             # handle operators that are both unary and binary, minimal BODMAS
             if op in ('+', '-'):
+                # simplify/adjust consecutive instances of these operators
                 undone = 0
                 while len(separated) > 1 and not separated[-1].strip():
                     undone += 1
                     separated.pop()
                 if op == '-' and undone % 2 != 0:
                     right_expr = op + right_expr
+                elif op == '+':
+                    while len(separated) > 1 and separated[-1].strip() in self.OP_CHARS:
+                        right_expr = separated.pop() + right_expr
+                # hanging op at end of left => unary + (strip) or - (push right)
                 left_val = separated[-1]
                 for dm_op in ('*', '%', '/', '**'):
                     bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
                     if len(bodmas) > 1 and not bodmas[-1].strip():
                         expr = op.join(separated) + op + right_expr
-                        right_expr = None
+                        if len(separated) > 1:
+                            separated.pop()
+                            right_expr = op.join((left_val, right_expr))
+                        else:
+                            separated = [op.join((left_val, right_expr))]
+                            right_expr = None
                         break
                 if right_expr is None:
                     continue
@@ -797,6 +846,8 @@ class JSInterpreter(object):
 
             def eval_method():
                 if (variable, member) == ('console', 'debug'):
+                    if Debugger.ENABLED:
+                        Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
                     return
                 types = {
                     'String': compat_str,

youtube_dl/utils.py

@@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
         """ tb, if given, is the original traceback (so that it can be printed out).
         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
         """
-
+        self.orig_msg = msg
         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
             expected = True
         if video_id is not None: